The Nightmare is the Most Useful Dream
Ernest Hartmann, M.D.
Tufts University School of Medicine; Newton-Wellesley Hospital, Boston
2
I will argue here that far from being a “failed dream” or an aberration of some sort, the
nightmare is in fact the paradigmatic dream. The nightmare presents us, in the clearest
form, what is present in all dreams. Thus, the nightmare is the most useful dream in our
attempt to study dreams in general. I will also discuss ways in which the nightmare can
be useful to the dreamer (the functions of the nightmare).
Towards a Definition of the Nightmare
The term nightmare has had a somewhat confused history, since it has been used in a
number of different senses. Thus, some well-known past works on nightmares clearly
refer to what we now call night-terrors. This is true of Ernest Jones’s classic work On the
Nightmare (1931) as well as portions of John Mack’s work Nightmares and Human
Conflict (1970). Sometimes nightmare has even been used to describe an object one is
afraid of (e.g., “the nightmare in my closet”), rather than a frightening dream.
As with many concepts and notions in psychology, there is still no perfect definition of
the nightmare. However, it is now accepted that the nightmare is a long frightening
dream, which awakens the sleeper; it usually occurs during the second half of the sleep
period and, if the sleeper happens to be in a sleep laboratory, it can be shown that the
awakening comes from a long REM period (Kryger, et al, 1994; Hartmann, 1984). By
this modern definition, the nightmare can be clearly differentiated from the night-terror,
which is a sudden arousal accompanied by a feeling of terror, occurring early during the
sleep period and generally accompanied either by no remembered content or a single
distressing image. Thus, the night-terror is not a dream and almost all reasonably
introspective persons can themselves tell the difference. In the sleep laboratory, the
night-terror arises from a sudden arousal out of deep slow-wave sleep (stage 3 or stage 4
sleep).
Although the nightmare can be clearly differentiated from the night-terror, there is no
clear demarcation between a nightmare and an anxiety dream. Both are dreams; we can
think of a continuum of frightening dreams and say that the nightmare refers to the more
frightening and terrifying end of the continuum. Also, at the nightmare end of the
continuum one is more likely to say one was awakened by the nightmare, rather than
simply remembering it.
Perhaps because of these confusions and problems of definition, the nightmare has
sometimes been set to one side, as though it were not really a dream; or it has been
considered a “failed dream,” (Kramer, 1991), in both cases suggesting that understanding
nightmares is some sort of separate issue not truly connected with understanding dreams.
The Nightmare as a Paradigm for All Dreams
My view is the opposite. I believe that the classic nightmare is the perfect way to begin
an understanding of dreams in general. For the majority of people a nightmare is
probably the most powerful, forceful dream they will ever experience. And the
nightmare in a traumatic or stressful situation often provides the simplest, clearest case of
3
what may happen more obscurely in other dreams. I am not speaking here of the chronic,
repetitive, post-traumatic nightmare. The repetitive PTSD subtype of nightmare could
indeed be thought of as a “failed dream,” since it does not evolve in any way, is very
disturbing to the dreamer, and has no function attributed to it. These repetitive PTSD
nightmares are quite rare. They can be of interest in their own right and have been
studied, but are not my subject here (see Hartmann, E., 1996). I refer to the far more
common nightmare which is not usually repetitive, but occurs frequently at times of
trauma or stress, and can occur in a whole series of gradually evolving shapes.
My associates and I have been studying for some years the dreams and nightmares that
occur after a traumatic event. A nightmare we have seen very commonly goes something
like this: “I am walking along a beach, when suddenly a huge tidal wave comes and
overwhelms me. I am swept away—I don’t know whether I’ll get to the surface before I
wake up.” This nightmare is important because we can see so clearly what is going on.
The traumatized person, who for instance has recently been attacked or has just escaped
from a fire, dreams of drowning, or almost drowning in a tidal wave. What is going on
here? Obviously the person is not dreaming about the actual events experienced, but
dreaming about the emotion associated with them. The dream contextualizes (pictures)
the emotional state of the dreamer. Typically, these dramatic nightmares are not
repetitive, but gradually evolve and change. These are especially useful objects of study,
at least for those of us (and I believe it is most of us) who believe that the emotional
concerns of the dreamer are somehow expressed in the dream and other mental activity
such as daydreams, fantasies, and stories. The reason that nightmares after trauma are
especially useful for the study of dreams is that the dominant emotion or emotional
concern is most evident in these cases. The first emotion is usually “I am terrified, I am
overwhelmed.” Later, there are often emotions such as grief or survivor guilt (“why was
I spared when the other person died or was injured?”).
Dreams and nightmares collected in numerous series from people who have undergone a
severe trauma frequently include dreams such as the following:
“A huge tidal wave is coming at me.”
“A house is burning and no one can get out.”
“A gang of evil men, Nazis maybe, are chasing me. I can’t get away.”
The person is dreaming the dominant emotion. The way we formulate it is that the
powerful dream image (tidal wave, etc.) contextualizes (supplies a picture-context for)
the underlying emotion — in this case fear or terror. Sometimes a slightly different
emotion which could be called vulnerability or helplessness is also involved producing
similar or slightly different dreams, such as:
4
“I dreamt about children, dolls — dolls and babies all drowning.”
“He skinned me and threw me in a heap with my sisters. I could feel the pain. I could
feel everything.”
Sometimes survivor guilt or other forms of guilt are dominant:
“A shell heads for us (just the way it really did) and blows up, but I can’t tell whether it’s
me or my buddy Jack who is blown up.”
“I let my children play by themselves and they get run over by a car.”
Thus, the suggestion is that the dream image contextualizes the dominant emotion or
emotional concern of the dreamer and that this can be seen most easily in the nightmares
after a single trauma, whereas in a less defined situation — a typical dream of a typical
student — there may be no single powerful emotion, rather a number of relatively mild
emotional concerns, and the dream is much less predictable (Hartmann, 1998).
Studies of the Contextualizing Image
The many examples, such as those mentioned, cannot be considered a research study,
since obviously we may have chosen examples that dramatically illustrate a thesis.
Therefore, we have recently attempted to approach these contextualizing images (CIs) in
a more quantitative manner.
First of all, a rating scale was devised for CIs in which a scorer, blind as to what sort of
dream material he or she is scoring, first decides whether there is a powerful, vivid,
detailed image, and gives it an intensity score from 0 (no image) to 3 (extremely
powerful). This is the CI score. Then the scorer tries to choose an emotion, out of a list
of 18 (see table 1), that might be contextualized by to this image. So far CI scores have
demonstrated good inter-rater reliability, from .60 - .85 on a variety of dream material.
When dreams and daydreams from college students were compared on a blind basis, the
CI score was shown to be higher (about twice as high) in dreams than in daydreams
(Hartmann, et al, 1998). Material obtained in subjects wearing a “nightcap” device,
enabling them to be awakened differentially during REM sleep, non-REM sleep, sleep
onset, or waking, was likewise scored on a blind basis for CIs. CI scores were
significantly higher in the material from REM sleep than from non-REM sleep, and the
material from non-REM sleep was significantly higher than material from waking
(Hartmann and Stickgold, 1999).
Returning to our initial interest in deciding whether CIs were more frequent and more
intense after trauma, we compared dream series of various lengths, obtained from 10
5
different persons who had experienced an acute trauma and had written down or told their
dreams in the period afterwards (a total of 451 dreams), with dreams reported by 306
college students. The CI scores overall were significantly higher in the trauma group
than in the student group (p < .001) (Hartmann, et al, 1998). In a related study,
examining the group of college students, we compared CI scores from those who on a
questionnaire reported no history whatever of sexual or physical abuse to those who
reported a history of any kind of abuse — physical or sexual, in childhood, adolescence,
or more recently. The dreams scored on a blind basis showed that the students who
reported any history of trauma had significantly higher CI scores than the other students
(Hartmann and Zborowski, 1999).
Thus, even though dreams are notoriously difficult to score reliably, we do have evidence
that CI scores are higher after trauma. Our hypothesis, obviously not easy to prove
conclusively, is that the same pattern of contextualizing emotional concerns seen above is
found in all dreams, although it is less evident. The pattern is most easily seen after a
traumatic event, but can also be found readily in stressful situations and in defined
situations such as pregnancy or in patients who suffer from one dominant emotion
(discussed in detail in Hartmann, 1998).
Thus I believe that the nightmare, far from being a failed or aberrant dream , is one of the
most important kinds of dream, and the one in which we can most easily observe a
process which probably occurs in all dreams. In this sense the nightmare is the most
useful of dreams.
6
An Adaptive Function of Nightmares?
Of course it is true that the usefulness outlined above is to the researcher or the
psychologist — to us as students of the dream — not necessarily to the dreamer
him/herself. Actually, I do believe that the nightmare may be useful to the dreamer as
well, and I will discuss this briefly here. It is taken up in great detail in Hartmann, 1998.
I discuss at length in my book the difficulty of proving or even attempting to prove
something like a function of dreaming, or a function of nightmares.
Without clear-cut research studies, but with a great deal of clinical experience, my
thinking is along the following lines. Starting again with the clearest case, an ordinary
adult who has experienced a single trauma, we often see a progression. First, there may
or may not be one or two dreams portraying the actual trauma. Then very quickly we
obtain various kinds of “tidal wave dreams” (nightmares). Successive dreams seem to
bring in more and more of the dreamer’s life, usually making use of emotional
connections. Thus, if the person is feeling primarily terrified or overwhelmed, other
situations connected with the same emotions will surface in dreams. If grief, guilt, etc.
are especially prominent, other situations involving these emotions will be found. I
believe that this whole developing series is of use to the dreamer. First of all, the dreams
reduce the intensity of the emotional distress by juxtaposing the current trauma with
various other events in the person’s life, making connections to other similar or not-sosimilar events, and so on. In addition to this immediate effect of “calming the storm,” I
believe the gradually evolving connections are useful in themselves. When trauma is
dreamt about (or thought about, fantasized about, worked on in therapy, etc.) it is no
longer uniquely distressing; it gradually becomes part of a fabric or a network.
Presumably the next time a similar trauma occurs, it will already have a context in which
it can be considered, rather than standing out as “the most awful thing that has ever
happened to anyone.” In this sense I believe that dreaming has an adaptive function, and
again that this adaptive function can be seen most readily when starting with the
developing nightmares following a trauma. Very similar views on an adaptive function
of dreaming have been voiced by a number of others, starting with very different data and
clinical experience. For instance, French and Fromm (1964), Greenberg and Pearlman
(1993), Breger (1971), Cartwright (1991), and Kramer (1993) have all proposed adaptive
functions of dreaming, though stated in somewhat different terms.
Other Uses of Nightmares
The above proposal of an adaptive function of dreams suggests a basic quasi-biological
function — something that is useful to the organism and increases its well-being and
chance of survival. In addition to this basic function, individuals have of course made
use of their nightmares in any number of different ways, especially in furthering their
self-knowledge with or without the help of psychotherapy. Artists have painted pictures
of their nightmares and other dreams, and writers have used a striking nightmare image
as a starting point for their stories. I have suggested elsewhere that the Christian idea of
hell probably derives from nightmares — certainly it is peopled with very typical
nightmare-creatures (Hartmann, 1984). It is clear that artists have been able to make use
7
of their nightmares in their art. This might be thought of as an occasional, rare use of
nightmares in a few creative people, but I believe it may occur more often than most
people believe. My evidence for this comes from studies I conducted some years ago of
persons who suffer from frequent nightmares — at least one per week, and had done so
for years. I studied 50 of these individuals in great detail (Hartmann, 1984), and obtained
data on at least 50 more persons with a similar history. After the people had described
nightmares to me in detail — and indeed the nightmares often sounded extremely
frightening and disturbing — I asked whether they had ever sought help or treatment for
the nightmares and whether they had even considered seeking treatment. Over half of the
group neither sought treatment nor considered treatment. When I asked them why, some
of them answered in terms of the use they made of nightmares in their artistic
productions. However, the most common answer was something like: “I don’t know.
The nightmares are so much a part of me. I think I must be making use of them in some
way, even though I don’t know exactly how.” So there are a surprising number of people
who believe that their nightmares are of use to them, even if only some are able to state
just what the use might be (art or self-knowledge).
Conclusion
In summary, I believe that nightmares are not aberrant or failed dreams. Rather, they can
be extremely useful and I have outlined several types of usefulness. I am most convinced
of the issue I raised first and discussed at length — that the nightmare is the most useful
dream because in the nightmare we can see most clearly the contextualization or
picturing of emotion, which I believe is a central aspect of all dreams.
8
Table 1 — Scoring Dreams for Contextualizing Images
Definition: A contextualizing image is a striking, arresting, or compelling image — not simply a story —
but an image which stands out by virtue of being especially powerful, vivid, bizarre, or detailed.
EMOTION LIST
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
fear, terror
helplessness, vulnerability, being trapped, being immobilized
anxiety, vigilance
guilt
grief, loss, sadness, abandonment, disappointment
despair, hopelessness (giving up)
anger, frustration
disturbing — cognitive dissonance, disorientation, weirdness
shame, inadequacy
disgust, repulsion
11.
12.
13.
14.
15.
16.
17.
18.
power, mastery supremacy
awe, wonder, mystery
happiness, joy, excitement
hope
peace, restfulness
longing
relief, safety
love (relationship)
If there is a second contextualizing image in a dream, score on a separate line.
Dream
ID#
1. CI?
(Y/N)
2. What is it?
3. Intensity
(rate 1-3)
4. What emotion?
5. Second emotion?
9
References
Cartwright, R. 1991. Dreams that work: The relation of dream incorporation to
adaptation to stressful events. Dreaming, 1:3-10.
French, T., Fromm, E.. 1964. Dream Interpretation. New York: Basic Books.
Greenberg, R., Pearlman, C. 1993. An integrated approach to dream theory:
Contributions from sleep research and clinical practice. In The Functions of
Dreaming. Moffit, A., Kramer, M., and Hoffman, R., eds. Albany: State
University of New York Press.
Hartmann, E. 1984. The Nightmare: The Psychology and Biology of Terrifying
Dreams. New York: Basic Books.
Hartmann, E. 1998. Dreams and Nightmares: The New Theory on the Origin and
Meaning of Dreams. New York: Plenum Press.
Hartmann E, Kunzendorf R, Rosen R, Grace N. 1998. Contextualizing Images in
Dreams and Daydreams. Sleep, 21S:279.
Hartmann E, Rosen R, Grace N. 1998. Contextualizing Images in Dreams: More
Frequent and More Intense After Trauma. Sleep, 21S:284.
Hartmann, E., Stickgold, R. 1999. Sleep, 22S, in press.
Hartmann, E., Zborowski, M. 1999. Sleep, 22S, in press.
Jones, E. 1931. On the Nightmare. London: Hogarth Press.
Kramer, M. 1991. The Nightmare: A Failure in Dream Function.
Dreaming, 6:277-285.
Kramer, M. 1993. The selective mood regulatory function of dreaming: an update and
10
revision. In The Functions of Dreaming. Moffit, A., Kramer, M., Hoffman, R.,
eds. Albany: State University of New York Press.
Kryger, M., et al. 1994. Principles and Practice of Sleep Medicine. Philadelphia:
Harcourt Brace Jovanovich.
Mack, J. 1970. Nightmares and Human Conflict. Boston:
Little, Brown and Company.